Mechanisms of Seed-To-Seed Interactions between Dominant Species in the Yangtze River Estuary under Saline Condition

: Plant community assembly is the central issue in community ecology. As plant traits differ in different life history stages, the form, intensity and mechanism of interspeciﬁc interactions may change with the ontogenetic process of plants. However, our understanding of interspeciﬁc interaction mechanisms during germination is still limited. Here, we conducted a laboratory germination experiment using ﬁve dominant species in Chongming Dongtan ( Spartina alterniﬂora , Scirpus mariqueter , Phragmites australis , Suaeda glauca and Tripolium vulgare ) to assess their germination performance in control (monoculture), allelopathy and mixture treatments. The results indicated that seeds could affect germination performance of neighbors through both allelopathy and salinity modiﬁcation. Salinity of the solution in Petri dishes after seed germination decreased signiﬁcantly in most species combinations in competition treatment, and was negatively correlated with the number of total germinated seeds. Seed leachate of invasive Spartina alterniﬂora signiﬁcantly accelerated the germination of two native halophytes Suaeda glauca and Tripolium vulgare , but not Scirpus mariqueter and Phragmites australis . The salt absorption by Spartina alterniﬂora seeds had inconsistent effects compared with that of its seed leachate. On the other hand, seed leachate of native Scirpus mariqueter and Phragmites australis signiﬁcantly slowed down the germination of invasive Spartina alterniﬂora . The effect of salinity modiﬁcation of Scirpus mariqueter on Spartina alterniﬂora was positive, whereas that of other species was neutral. Considering seed-to-seed interactions is an important perspective to understand the underlying mechanisms of community dynamics, species diversity maintenance and invasion of alien species, and can improve the effectiveness in the management of invaded coastal wetlands.


Introduction
The spatial distribution pattern of different plants and the underlying mechanism of species coexistence (i.e., plant community assembly mechanisms) have been central issues in community ecology for decades [1,2]. Interspecific interactions are thought of as the driving force of plant community assembly, especially in the later successional stages [3,4]. Positive interactions often occur in stressful environment, and plants can facilitate the survival, growth and reproduction of their neighbors via habitat modification [5]. In relatively benign environment, competition is more common, which can negatively affect the performance of adjacent plant species through resource contention, allelopathy and other mechanisms [6].
Previous studies on interspecific interactions mainly focused on established seedlings and ramets [7][8][9]. However, as plant traits differ in different life history stages, the form, intensity and mechanism of interspecific interactions may change with the ontogenetic salinity of surface soil in Chongming Dongtan usually varies from 2‰ to 15‰, and can exceed 20‰ during salt water intrusion [41].
Major native plant communities in Chongming Dongtan are sedge meadows dominated by Scirpus mariqueter (Sm) in middle to low marshes and grass marshes dominated by Phragmites australis (Pa) at higher elevations. Scirpus mariqueter is a perennial rhizomatous species mainly distributed in the Yangtze River estuary and the Hangzhou Bay. It can reproduce both via seeds and via corms and rhizomes. As a pioneer species, it can colonize stressful habitats where other plants cannot survive [36]. Phragmites australis (Pa) is a perennial species that commonly distributes in inland or coastal wetlands [42]. It mainly relies on vegetative reproduction through rhizomes in mature populations, whereas seedlings play important roles in colonization of bare patches. In high marsh, Suaeda glauca (Sg) and Tripolium vulgare (Tv) often occurs in hypersaline bare patches. Suaeda glauca is an annual halophyte which often forms large scale communities around saline-alkaline lakes. It has high salt tolerance and can also grow in intertidal zones frequently affected by tides [43]. Tripolium vulgare is also an annual forb which often grows in alkaline lake wetlands with soil pH ranges from 8.10 to 9.15. It can invade Phragmites australis community under saline and dry conditions [44]. After the introduction of Spartina alterniflora (Sa), this invasive species thrived in Chongming Dongtan from middle to high marsh and formed Scirpus mariqueter-Spartina alterniflora and Spartina alterniflora-Phragmites australis mixtures [13].

Experimental Design
The seeds of Spartina alterniflora (Sa), Scirpus mariqueter (Sm), Phragmites australis (Pa), Suaeda glauca (Sg) and Tripolium vulgare (Tv) were collected in the pure stands of each species in Chongming Dongtan in early October, 2017. The collected seeds were stored in a refrigerator at 4 • C. Spartina alterniflora and Scirpus mariqueter seeds were immersed in fresh water and the others were kept in dry condition to maintain seed vigor [13]. The germination experiment began on 13 March 2018. All the seeds were washed with distilled water for 3~5 min three times and then placed in a Petri dish (9 cm in diameter) with two pieces of filter paper on the bottom. The Petri dishes were placed in a light incubator (PGX-600B) for germination, with the setting of 30 • C and 100% light intensity (12,000 lux) in the daytime (from 08:00 to 20:00) and 18 • C and 0% light intensity in the night (from 20:00 to 08:00).
For preparation of seed leachate, 50 g of Spartina alterniflora, Scirpus mariqueter, Phragmites australis, Suaeda glauca and Tripolium vulgare seeds were placed in a 500 mL conical flask, respectively. A total of 400 mL of distilled water was added into each flask. The flasks were sealed and vibrated in a shaker at room temperature for 48 h to extract the maximum concentration of allelochemicals from seeds [45]. After that, NaCl was added into the flasks to reach the salinity of 1%, which is close to the condition of their natural habitats in Chongming Dongtan.
To highlight the effects of invasive plant on native species during germination, this study focused on the interactions between the seeds of invasive species and native species rather than those between different native species. Hence, there were four species combinations: Sa + Sm, Sa + Pa, Sa + Sg and Sa + Tv. We set three different treatments to all species combinations: control (target species seeds in monoculture), allelopathy (target species seeds + competitor seeds leachate) and competition (target species seeds + competitor seeds in mixture). In the control treatment, 20 seeds of a single species were evenly placed in a Petri dish and 6 mL of 1% NaCl solution was added. In the allelopathy treatment, 20 seeds of each species were evenly placed in a Petri dish with 6 mL seed leachate of its competitor seeds. Spartina alterniflora seeds were treated by leachate of Scirpus mariqueter, Phragmites australis, Suaeda glauca and Tripolium vulgare, respectively. Similarly, four native species seeds were treated by Spartina alterniflora leachate, respectively. In the competition treatment (Sa + Sm, Sa + Pa, Sa + Sg and Sa + Tv), 20 seeds of each competing species were placed in mixture in each Petri dish and 6 mL of 1% NaCl solution was added. Our experiment followed an additive design to focus on interspecific interactions rather than intraspecific interactions among these species. The two density levels (20 seeds per dish or 3144 seeds m −2 , 40 seeds per dish or 6288 seeds m −2 ) were set according to the soil seed bank density in Chongming Dongtan (250~8500 seeds m −2 for Scirpus mariqueter and 500~6900 seeds m −2 for Spartina alterniflora) [36]. There were 17 treatments in all (5 control + 8 allelopathy + 4 competition) and each treatment was replicated four times, making a total of 68 dishes.
Distilled water was added into each Petri dish every day to maintain a shallow layer of water. The number of germinated seeds (where root tips had protruded from the seed coat) of each species was counted and recorded every day, and then, all the Petri dishes were randomly rotated to change its position in the incubator until no seeds germinated within a week. The germination experiment lasted for about a month (13 March to 16 April). After germination, the solution in each Petri dish was diluted to 30 mL, and the salinity was determined by a portable salinometer (SANXIN 5052).

Data Analysis
Time lag was calculated as the time between the beginning of the experiment and the first germination of each species. Germination period was the time between the first germination and the last germination of each species. Germination percentage was calculated as the proportion of total germinated seeds of each species. Germination index (GI) was calculated as follows [46]: Mean time to germination (MTG) was calculated as follows [47]: where Gi is the number of germinating seeds on the ith day and Di is the number of days after the beginning of the germination experiment.
One-way ANOVA (post hoc Tukey's HSD test) was used to examine the differences of germination traits (time lag, germination period and germination index) of each species in the control treatment. For each competing species in each species combination, one-way ANOVA (post hoc Tukey's HSD test) was used to examine the effects of different treatments on germination percentage, MTG and the final salinity of solution in Petri dishes after germination. The relationships between final salinity and the number of total germinated seeds were analyzed using linear regression for each species combination. The original germination data were log transformed if necessary to meet the assumption of homogeneity of variance. The significance level was set to 0.05 (p < 0.05). All the statistical analyses were carried out by STATISTICA 13.5 (TIBCO software).

Germination Traits of Dominant Species
Tripolium vulgare had the highest time lag among the species studied, followed by Scirpus mariqueter. Spartina alterniflora and Suaeda glauca had a significantly lower time lag than Tripolium vulgare and Scirpus mariqueter, while Phragmites australis fell in between ( Figure 1a). Tripolium vulgare had a significantly higher germination period than Phragmites australis, while Spartina alterniflora, Suaeda glauca and Scirpus mariqueter had an intermediate germination period (Figure 1b). The germination index of Suaeda glauca and Spartina alterniflora was significantly higher than that of Phragmites australis, Scirpus mariqueter and Tripolium vulgare (Figure 1c).

Effects of Interspecific Interaction on Germination
In species combination of Sa + Sm, the germination percentage of both species was significantly promoted in competition treatment than in control and allelopathy treatments (Table 1, Figure 2a). The germination percentage of Phragmites australis seeds was also significantly affected by interspecific interaction ( Table 1). The allelopathy treatment inhibited the germination of Phragmites australis seeds, resulting in a significantly lower germination percentage than in the competition treatment, which had a promotion effect  The germination speed of most species (indicated by mean time to germination) was significantly affected by interspecific interaction treatment except for Phragmites australis seeds ( Table 2). The germination of Spartina alterniflora seeds was the fastest (with the lowest mean time to germination) in control treatment and was significantly slowed down in allelopathy treatment with Scirpus mariqueter and Phragmites australis leachate (Figure 3a,b) and in competition treatment with Suaeda glauca and Tripolium vulgare seeds (Figure 3c,d). Competition treatment with Spartina alterniflora seeds significantly slowed down the germination of Scirpus mariqueter (Figure 3a), whereas allelopathy treatment with Spartina alterniflora leachate significantly promoted the germination speed of Suaeda glauca and Tripolium vulgare (Figure 3c,d).

Salinity Modification during Germination
The final salinity of the solution in Petri dishes after germination differed significantly among different interspecific interaction treatments for Spartina alterniflora in species combinations of Sa + Sm and Sa + Sg, and for Scirpus mariqueter, Phragmites australis and Tripolium vulgare (Table 3). In most species combinations, final salinity was significantly lower in competition treatment than in control and allelopathy treatments, except for Spartina alterniflora in species combinations of Sa + Pa and Sa + Tv, and for Suaeda glauca in species combination of Sa + Sg (Figure 4). Table 3. Results of one-way ANOVA, testing the effects of interspecific interaction treatment on the final salinity of the solution in Petri dishes after seed germination in different species combinations. The abbreviations in brackets after species combinations denote the target species considered. Significant p values are in bold. In all species combinations, the final salinity of the solution in Petri dishes after germination significantly decreased with an increasing total germinated seed number ( Figure 5).

Discussion
Seed germination is an irreversible process and the weakest stage in the life cycle of plants [48], which plays important roles in maintaining plant populations and communities. The findings in this and previous studies [41] suggested that the interaction between seeds is more complicated than we realized before. Our study indicated that the leachate of competitor seeds and the entire competitor seeds had inconsistent effects on germination of target species. This, together with the reduction of solution salinity in Petri dishes, implied that seeds could affect germination performance of neighbors through both allelochemical release and microenvironment modification.

Effects of Allelopathy on Seed Germination
In a natural environment, soil seed banks usually consist of seeds of different species, which makes seed-to-seed interactions ubiquitous. Adult plants can affect neighbors in a number of ways (e.g., resource competition and direct interference), whereas seeds interact with each other primarily through allelopathy [49]. In addition, the nature and magnitude of interspecific interactions among seeds usually vary with different types and concentration of allelochemicals [50].
In our experiment, the leachate of invasive Spartina alterniflora seeds slightly reduced (but not significantly) both the germination percentage and germination speed of native Scirpus mariqueter and Phragmites australis (Figures 2a,b and 3a,b). That is to say, only a small part of seeds of these species germinated slowly, while the others failed. This effect was probably caused by allelochemicals released from Spartina alterniflora seeds. As the shape and anatomical structure of seeds (e.g., seed coat thickness, wax content and embryo position) play important roles in seed response to external conditions [48,51], seed leachate will have different effects on different species. That might be the reason why the germination process of Suaeda glauca and Tripolium vulgare was significantly accelerated (Figure 3c,d).
The leachate of native Scirpus mariqueter and Phragmites australis seeds altered the microenvironment and significantly reduced the germination speed of invasive Spartina alterniflora seeds (Figure 3a,b). That means that more Spartina alterniflora seeds (especially those with poor germination ability) were activated to germination. These asymmetric effects between invasive and native seeds may be due to their substantial mass difference [35,[52][53][54]. During germination, all the seeds may absorb allelochemicals from surrounding leachate, but accumulate them in different concentration, and the allelopathic effects of these chemicals usually depend on the concentration [31,50]. We are not clear what kind of allelochemical was released in the seed leachate of these species and how the interaction effects vary with their concentration, which needs further validation.

Effects of Salinity Modification on Seed Germination
During germination, seeds can absorb ions from the surrounding solution, which allow them to regulate osmotic potential [55,56]. We found a decrease in salinity after germination in most species combinations in competition treatment (Figure 4), which indicated the importance of microenvironment modification in mediating seed-to-seed interactions. Under the additive design, seed density in competition treatment was twice that of other treatments, which might be the main cause of decreased salinity in mixture. However, the change of total density is also a part of the influence of competing species. Future research following a response surface design can help us understand the effects of both intraspecific and interspecific interactions in different species proportion and total density levels. Notably, seeds in the mixture treatment are affected by allelopathy and microenvironment modification simultaneously, and we are not able to separate them from each other.
In our experiment, invasive Spartina alterniflora seeds had inconsistent effects compared with seed leachate and promoted the germination percentage of Scirpus mariqueter seeds (Figure 2a,b). Meanwhile, the germination speed of Scirpus mariqueter was significantly slowed down (Figure 3a), whereas that of Phragmites australis was not affected (Figure 3b). This can be explained by the positive effects of decrease in solution salinity in mixture. Scirpus mariqueter seeds are salt sensitive [13]. When the seeds of this native species germinate alone in 1% NaCl treatment, they will suffer from high osmotic and ion-toxicity stress, and only a small proportion of seeds successfully germinate. As Spartina alterniflora seeds are more salt tolerant, a great deal of salt ions will be absorbed during germination when this invasive plant is present, which led to a significant decrease in solution salinity. The positive effect of salinity modification by Spartina alterniflora on native species seemed to be stronger than the negative effect of its seed leachate, as the germination percentage of Scirpus mariqueter seeds was the highest in the mixture (Figure 2a). Conversely, halophyte Suaeda glauca and Tripolium vulgare seeds are much more salt tolerant, and require a certain concentration of salt during germination. Hence, alleviation of salt stress by competing Spartina alterniflora seeds in mixture had minor effects, which did not significantly affect the germination percentage (Figure 2c,d) and germination speed of these two species (Figure 3c,d).
For Spartina alterniflora, salt-sensitive Scirpus mariqueter seeds significantly promoted its germination percentage (Figure 2a), whereas the seeds of halophyte Suaeda glauca and Tripolium vulgare significantly delayed its germination process (Figure 3c,d). The net effects of microenvironment modification (i.e., regulation of salinity) on this invasive plant differed in different combinations with native species, which was also related to their salt preference. Despite of different species combinations and treatments, the extent of solution salinity decrease in Petri dishes after seed germination is all significantly correlated with the number of total germinated seeds ( Figure 5), indicating that the salinity modification effect by competing seeds was mainly due to the change of total density level rather than the biological attributes of competitors. Surprisingly, the Sa + Sg combination had the lowest efficiency of salt absorption per single seed (Figure 5c), even though these two species were the most salt tolerant ones among the five species used in our study. Whether there is another interaction mechanism during seed germination needs further investigation.

Implications for Community Assembly
Interspecific interactions play important roles in species coexistence and biodiversity. As seeds of different species coexist universally in the field, they may interact with each other through leachate or the modification of microenvironment in the soil. Seeds that germinate earlier usually have an advantage over those germinate later in competition [57]. For invasive plants, high seed yield, high germination percentage and short germination time are important traits which facilitate their invasion into native plant communities [58]. However, sometimes, rapid germination causes seeds to be exposed to a stressful environment, and is bad for subsequent plant survival [59]. Previous studies found that there exists a special mechanism for seeds to "sense" the surrounding environment via allelochemicals, which allow them to assess the optimal germination timing for more competitive advantages [57,60,61]. Among the five species used in our study, Phragmites australis had the shortest germination period (Figure 1b), and hence, the lowest potential of germination timing regulation. On the contrary, Tripolium vulgare had the longest germination period (Figure 1b), which may improve its adaptability to changing environments. It should be noted that due to the difference in germination timing, seeds that germinate later will not only affected by neighboring seeds, but also by seedlings from early germinated seeds [62]. Future research on the interaction between seeds and seedlings will contribute to our understanding of the role of the germination process in community assembly.
Our study found inconsistent effects of seed leachate and salinity modification on germination of both invasive and native species in Chongming Dongtan. On the whole, Spartina alterniflora had more positive effects on Scirpus mariqueter and Phragmites australis seeds during germination than on halophyte Suaeda glauca and Tripolium vulgare, and vice versa. Although invasive Spartina alterniflora promoted the germination percentage of Scirpus mariqueter seeds, it may still gain competitive advantage over native Scirpus mariqueter, Tripolium vulgare and Phragmites australis due to its short time lag (Figure 1a) and high germination index (Figure 1c). Besides, the strong competitive ability of Spartina alterniflora in subsequent life stages would promote its invasion into these native plant communities and had caused a series of negative ecosystem impacts [37,63,64]. The competitive mechanisms between invasive Spartina alterniflora and native species during germination stage were quite different from those between adult individuals, which highlighted the importance of including regeneration into the framework of community assembly.
In Chongming Dongtan, perennial clonal plants Spartina alterniflora, Phragmites australis and Scirpus mariqueter are heavily dependent on vegetative reproduction in mature patches, whereas seed dispersal and germination play important roles in new habitat colonization [13]. Therefore, eradication of both the aboveground and belowground part of Spartina alterniflora before its florescence is critical for control of this invasive species. After the elimination of Spartina alterniflora, transplantation of Phragmites australis should be implemented due to its relatively low seed vigor in saline water, whereas Scirpus mariqueter has the potential to recover from its persistent seed bank [65]. Artificial reduction of sediment salinity will be an effective measure to facilitate the recovery of these native species, but the cost is too high under current conditions. On the other hand, seed dispersal and germination are crucial for the recruitment of annual Suaeda glauca and Tripolium vulgare, but they are less affected by invasive Spartina alterniflora in their hypersaline habitats. Seed addition of these halophytes may also be beneficial to community biodiversity, as there is intransitive competition among the dominant species in Chongming Dongtan during germination stage [66].
Seed-to-seed interactions are an important perspective to understand the underlying mechanisms of community dynamics, species diversity maintenance and invasion of alien species and can improve the effectiveness in the management of invaded coastal wetlands. However, it should be noted that due to the frequent scouring of tide in the field, the concentration of allelochemicals in the soil will be much lower than in laboratory treatment, and seeds will have minor effects of salinity modification. We need to be very careful to extrapolate our laboratory results to natural ecosystems.